Hi TH1::Merge is unfortunately not able to merge histograms filled with labels. This functionality remains to be implemented. For the time being, you must loop on all your files, loop on all histogram bins, getting the bin label and filling the original histogram. Rene Brun On Fri, 1 Aug 2003, Weijiang Dong wrote: > Hi, > I want to plot the histogram of number of emc points for each runid > to check bad runs. > The runid is an integer. However, the runid has 7 digits. For a > whole year's experiment, there would be only like a hundred of different > runids. To make it worse, the last 6 digits of the integer runid can be > anything. So, by using TH1F with 1000000 bins, each bin will be slim as a > line on the plot. > I found I can convert the integer runid to string and fill the > histogram with the string and a weight(hlabel1.C and hlabel2.C in root > tutorial). It worked fine. I submitted hundreds of jobs and each > job generated a root file containing a histogram like that. But finally I > want to merge these root files. > I ever used the example macro hadd.C to do it. But root > complained histograms in each root file have different bin size. It's > true. Because not every runid would present for a job, so not every > runid string would fill the histogram. Later I found hadd.C used > TH1->Add(TH1 *) method and the "Add" method requires each histogram should > have the same bin size. Then I changed the hadd.C macro a little > bit. This time I used TH1->Merge(TList *). Root didn't complain anything > this time, but the result is strange and not correct. > Am I doing right or I missed something? Is there any other simple way to > do this? > Thanks. > Weijiang > >
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:14 MET